[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [MiNT] on Mega ST



Michael Bernstein wrote:

> Hi Helmut,
>
> > > - MiNT did not find the C:\MINT folder
> >
> > That's true, should be fixed. It only tries <sysdir>/mint.cnf.
> > Additionally it could try <sysdir>../mint.cnf. Would that be ok?
>
> If <sysdir> is C:\MINT\1-17-0 not. I could not create a folder with
> - in the name with a desktop from a TOS up to 3.xx So MiNT should
> use also the searchpath <bootdrive>:\MINT

My theory was that TOS would drop the last path-item (1-17-0). But as I
said in the other post, the code fot \MINT\ has been removed (ozk?).

I no one objects I will reactivate it.

> > > - ext2 driver makes the system crash if the system access the filesystem
> >
> > That is a severe bug. What if there exists an ext2-fs?
>
> I did nit create a ext2 partition. I can do it on my Mega ST. But i

Maybe there's a bug that leads the PC into data-segment or similar. I
just tried in aranym-mmu without an ext2-partition and there was no error.

> think, this will not avoid the crash. Because i get a invalid
> instruction exception, i assume the ext2 driver was a only 030
> software.

The only asm is this (bitmap.h):

INLINE long
ext2_set_bit (long nr, void *vaddr)
{
  char retval;

  __asm__ __volatile__
  ("\
    bfset %2@{%1:#1};     \
    sne %0 "      \
    : "=r" (retval)     \
    : "d" (nr^7), "a" (vaddr) \
  );

  return (long) retval;
}

INLINE long
ext2_clear_bit (long nr, void *vaddr)
{
  char retval;

  __asm__ __volatile__
  (" \
    bfclr %2@{%1:#1}; \
    sne %0 "
    : "=d" (retval)
    : "d" (nr^7), "a" (vaddr)
  );

  return (long) retval;
}

INLINE long
ext2_test_bit (long nr, const void *addr)
{
  char retval;

  __asm__ __volatile__
  ("\
    bftst %2@{%1:#1}; \
    sne %0 "
     : "=d" (retval)
    : "d" (nr^7), "a" (addr)
  );

  return (long) retval;
}


Maybe someone with more knowledge can tell.


-Helmut